i'm working on a SPI NAND Flash applet for #GlasgowInterfaceExplorer. i would like to add the ability to read (and write) data dumps... however, unlike with SPI NOR flashes where the data is located in an flat address space, SPI NAND devices are structured as pages (each consisting of data + spare + ECC)

how should i format these dumps? is there a standard format? is "use msgpack with a simple custom schema" good enough, considering msgspec promises fast (de)serialization for Python and other languages?

if you have a #GlasgowInterfaceExplorer, please try out the new memory-25q applet and let me know if it works for you!

https://glasgow-embedded.org/latest/applets/memory/25q.html

so!! i am excited!!! to have finally finished the complete reimplementation of the #GlasgowInterfaceExplorer memory-25x applet for managing SPI NOR flashes. it is called memory-25q and it took an enormous amount of effort, because i have decided to Build It Properly

want to jump to the docs (there are a lot of docs, including on the fundamentals of (Q)SPI flashes) or read the code? here we go:

• pull request introducing memory-25q
• quad SPI controller API
• framework for managing QSPI-style memories
• the CLI of the new memory-25q applet
• and, finally, the API of the new memory-25q applet

now, why did i do that? two reasons. memory-25x is one of the first applets i made, ~7 years ago, and i had no idea what kind of UI i should be building (yet). to make it worse, i thought that SPI NOR flashes were "easy", you could "just send a few bytes and that's basically it".

nothing could be further from truth. first off, SPI NOR flashes don't really exist—there is no spec, no standard organization that can say "no, your thing is not compliant", no order to any of this. every vendor does whatever they want, and then every other year JEDEC writes down all of the unhinged shit they did. here is the list of six incompatible methods to turn a single bit on or off, as a warmup

second, SPI flashes have an absolutely absurd diversity of framings. you cannot even express it without building a meta-framework for abstracting over all the ways people have come up to squeeze 8 bits into 2 or 4 wires. then on top of it you have to manage a bunch of global state that affects framing in subtle or sometimes really fundamental ways, without having any way to find out that you've made an error besides "you compare the actual data with the expected data (or its checksum) and it is not equal"

anyway, the new applet should be excellent at any daily task and at least okay at >90% of the exotic ones. also it's easily generalized for the (completely incompatible on the wire) QSPI NAND 25N series, octal or DTR variants, etc

So #HackerBoxes dropped of number 124 today "Bus Driver", and I'm looking forward to using it. But I also realized that it has the basis for a good workshop on exploring hardware interfaces which I use #GlasgowInterfaceExplorer for.

https://hackerboxes.com/collections/past-hackerboxes/products/hackerbox-0124-bus-driver

today's project for #GlasgowInterfaceExplorer: reworking the memory-25x applet

this is a very old applet, one of the first introduced. it works mostly fine, but it was designed and implemented before I really knew how to make good applet UIs, and long before I added QSPI support, so it's not using all of the available bandwidth & it requires specifying a bunch of geometry parameters by hand even when it is trivially extracted from the flash itself

I am planning to redesign the Flash command engine to support not just 1-1-1 (normal SPI) commands, but every mode including 1-1-2 and 1-1-4 (dual/quad data) and 1-2-2 and 1-4-4 (dual/quad address and data). to do this I will implement an abstraction for a (Q)SPI command set, which can then be extended and specialized by an individual flash

this will also handle devices like Winbond's and Micron's QSPI NAND devices (W25N series and MT29F series, respectively).